Operational Decision Architecture defines where AI may participate and where human authority must hold.

Operational Decision Architecture defines how an operation decides: which decisions exist, which roles hold authority, which inputs matter, which conditions trigger escalation, where AI may participate, where human judgment must lead, and what must be captured when the decision is made.

In aviation, that includes decisions around dispatch, release, recovery, crew constraints, maintenance coordination, customer pressure, irregular operations, and operational control. The same discipline applies to other regulated environments where decisions must remain accountable even when AI tools participate.

ODA is not a dashboard, chatbot, or vendor policy engine. It is the operator-side architecture that lets an organization govern AI participation across tools, accounts, vendors, and workflows.

Operational Decision Provenance, or ODP, is the record of how a specific decision formed: context, inputs, authority, rationale, AI participation, action, and outcome.

The relationship is simple: Operational Decision Architecture defines the decision class. Operational Decision Provenance captures the decision instance. Without ODA, the operation may log activity but still lose the reasoning, authority boundary, and judgment that made the decision defensible.

That distinction matters as AI moves from productivity tools into operational workflows. Vendors govern their tools. Operators govern the decisions those tools touch.